From the Mountains to the Machine

Every transformative technology company begins with someone who becomes tired of watching talented people spend their time on repetitive work.

For Muboriz Muborakshoev, that observation followed an unusual journey — from the mountains of Central Asia to London's financial sector and ultimately Dubai.

Before Astel Ventures, Muboriz worked across venture capital and corporate finance, experiencing firsthand the friction of manual due diligence, fragmented investor networks and outdated workflows.

Dubai provided a different operating environment: ambitious, fast-moving and increasingly comfortable with technology-led financial innovation. Astel Ventures emerged from that intersection.

The four-day problem

What happens when artificial intelligence compresses days of investment work into hours?

Astel's proposition is not simply to add another AI interface to financial services. It is to attack the operational friction surrounding deal execution.

For advisory firms, corporate-finance teams, investment banks and founders, research that traditionally requires analysts to move manually between documents, databases and networks can increasingly be orchestrated through AI.

The objective is not removing judgement from investment decisions. It is giving human judgement better information, faster.

Due diligence. Old way: reviewing 40–50 documents over 3–4 days. Astel approach: comprehensive investment analysis generated within approximately 3–4 hours.

Investor mapping. Old way: weeks of manual networking and cold outreach. Astel approach: path-of-least-resistance investor mapping in hours.

Team scaling. Old way: expanding analytical teams as workload increases. Astel approach: scaling output without proportional headcount growth.

Tokenmaxxing

When measuring AI adoption accidentally makes people less productive.

As companies race to demonstrate AI adoption, Muboriz has noticed an unintended management problem.

Some organisations are beginning to measure AI engagement through token consumption — effectively tracking how much artificial intelligence employees use.

The metric appears logical. But once consumption becomes a KPI, behaviour changes. Employees can be incentivised to consume more tokens simply to demonstrate that they are “AI-forward,” regardless of whether those tokens create better outcomes.

“When consumption becomes the target, consumption stops being a useful measure of productivity.”

Build what customers want. Not what's in your head.

Muboriz argues that one of the most dangerous environments for a founder is the founder's own imagination. A product can contain ten features its creator loves while customers consistently return for only one.

His response is deliberately analogue: conversations. Before over-engineering a product, founders should speak directly with dozens of potential users, observe how they actually work and listen before building. Even after product-market fit begins to emerge, the conversation should continue.

Hire slow. Fire fast. Conviction matters more than simply filling a position.

Give real ownership. Accountability becomes powerful when people can see their direct impact on what they are building.

What's next: the compute wars

As AI adoption accelerates, Muboriz expects compute capacity, infrastructure and token economics to become increasingly important competitive constraints.

The next battle may therefore be fought not only over who builds the strongest models, but over who can deliver intelligence economically and at scale.

Editorial opinion

The first wave of artificial intelligence was about capability. The next will be about execution. Astel Ventures sits inside that transition. Muboriz Muborakshoev is not betting on AI replacing the investment professional; he is betting that the investment professional who learns to command AI will operate at a fundamentally different speed.

That distinction matters. Compressing days of diligence into hours is not simply automation. It changes how quickly capital can understand an opportunity, challenge an assumption and make a decision.

But perhaps Astel's more interesting observation is what happens when companies mistake AI activity for AI productivity. “Tokenmaxxing” is a warning for the coming enterprise-AI era: measuring consumption is easy; measuring intelligence gained from that consumption is much harder.

Muboriz's philosophy ultimately returns to something decidedly human — customers, judgement, ownership and resilience. The technology may become superhuman. The conviction behind it still has to be human.

The machines are getting faster. The advantage will belong to those who know what to ask of them.

“AI-enabled is no longer a career option. It is becoming a survival mechanism.”